Surrogate-Based Noise Prediction in Actuator Line Method for HAWTs

Authors

Lorenzo Tieghi - Sapienza University of Rome, Italy
Alessandro Corsini - Sapienza University of Rome, Italy
Giovanni Delibra - Sapienza University of Rome, Italy
Filippo De Girolamo* - Sapienza University of Rome, Italy

Abstract

Wind turbines are considered to play a major role in European Green Deal for Clean Energy Transitions. Installation on- and off- shore have steeply increased the production in the last 10 years. Noise poses a significant role among open issues in both kind of installations, as it determines the possibility of installations near inhabited places and the possible detrimental effects on wildlife when off-shore.However noise prediction of wind turbines poses a series of challenges as the requirements for a proper CFD and CAA analysis are fare beyond the computational power available to industry and experimental setups still lack a standardized ISO procedure due to the limitations in controlling environmental conditions during tests and the problems of using wind tunnel measurements due to lack of reliable scaling laws.In this paper we discuss a numerical methodology to predict wind turbine sound pressure level based on surrogate models and URANS computations carried out with actuator line model (ALM). ALM computations have become dominant in this field due to the possibility of reducing the computational requirements of a full rotor solution. They were found able to predict turbine performance and reproduce a realistic turbulent wake. However, the lack of a proper solution of the boundary layer does not allow for full CAA analysis as major contributions to turbine noise are neglected and therefore there is need for surrogate models to account for these limitations.In this work an isolated wind turbine was simulated in a CFD environment based on OpenFOAM v1812 and an open-source library to implement the Actuator Line method of [1]. The selected turbine is the Vestas NM80 [2] and the accuracy of this coupled technique is thoroughly tested and compared with experimental data of the measurement campaign conducted in the DANAERO project [3].Computations entail the effects of the atmospheric boundary layer and conditions corresponding to different operating points of the turbine.The surrogate model for noise prediction is based on Synthetic/Surrogate Acoustic Models (SAM) of Amiet’s and Brooks-Pope-Marcolini (BPM) and allows to not model blade motion and to extract noise data from the fluid dynamic field in post-processing, resulting in an ordinary URANS.[1] Bachant, P., Goude, A., Wosnik, M., turbinesFoam/turbinesFoam Version v0.0.8, Zenodo, http://doi.org/ 10.5281/zenodo.1210366, March 2018.[2] Madsen, H.A., Bak, C., Paulsen, U.S., Gaunaa, M., Fuglsang, P., Romblad, J., Olesen, N.A., Enevoldsen, P.B., Laursen, J., & Jensen, L.E. The DAN-AERO MW experiments final report. Risø-R-1726(EN). Roskilde, Denmark: Risø National Laboratory; 2010.[3] Troldborg, N., Bak, C., Madsen, H.A., & Skrzypinski, W. R. (2013). DANAERO MW: Final Report. DTU Wind Energy. DTU Wind Energy E No. 0027(EN)